lcalcote

A Multiplayer Kubernetes Designer - Explore the Cloud Native Computing Foundation's graduated, incubation, and sandbox projects as well as many other popular open source projects. Use Meshery Playground to explore a new way of DevOps - visual and collaborative configuration management for your infrastructure.

In this blog post, we will be discussing a new feature in Kubernetes called "Validating Admission Policies" which is currently in alpha stage. This feature allows administrators to define custom validation rules for resources being created in the cluster and enforce those rules using admission controllers.

Meshery, the open source cloud-native manager, has just reached a huge milestone of 1,000 contributors! This is an incredible achievement, and we would like to extend a huge thank you to all of our users and contributors for their support. It is thanks to your dedication and hard work that Meshery as a project and community has grown so quickly. The project has garnered over 2,000 GitHub stars to date and the community has become the #1 most popular Linux Foundation Internship (LFX internship).

Not all, but some other vendor solutions autoinstrument popular languages, frameworks, runtimes as well. Many, many customers leverage these capabilities happily, relying on those engineered solutions (akin to the reliance others place on the engineering put into OC).

It is a good note to make. How does this bear weight on the question above?

As I help lead the Istio User Experience Working Group, I can confirm that there's a cost to telemetry gathered and sent to Mixer and that tuning the type and amount of telemetry sent can can a dramatic impact on the overhead involved in using the service mesh.

In general, the question of overhead is concern that most adopters of a service mesh ask. Akin to deploying an application performance monitoring agent to your server or an SNMP poller to your network, there's a price to be paid for the ability to observe and control your workloads. Illuminating this cost, understanding and controlling this cost is the goal of an open source project, Meshery (https://layer5.io/meshery). Meshery acts as a performance benchmark for multi-mesh performance comparison (an apples-to-apples comparison across different service meshes).

Note that it's not necessary for each sidecar proxy call to validate every network call with Mixer. Many of these validations are locally cached in the sidecar proxy and do not involve Mixer.

One way of interpreting this simile is to say that as just as one technology has subsumed (become the focal point) another technology, so does a service mesh as a layer on top of a container orchestrator. A fifth layer as described in The Enterprise Path to Service Mesh Architectures --> https://layer5.io/books/the-enterprise-path-to-service-mesh

Good summary.

50 free checks at 5 min intervals. Hmm.

The service mesh landscape will be updated soon - https://layer5.io/landscape

Users shouldn't be confused as they are able to continue using docker as they have in the past. The Moby project doesn't affect their consumption and use of docker.

@ncresswell, more create and update capabilities would be great (e.g. the ability to create a network, connect a container, etc.)

"That's because ELBs forward requests to backend instances without modifying the request headers and the client IP is not sent in case of TCP load balancing." - - not sure this is quite what the author meant?

This is true. The team packed a lot into this new codebase. Many miles to go, however.

Are next steps for multi-zone support inclusive of these capabilities?

1) HA for master components 2) Multi-zone support for PersistentVolumes 3) Federated master components governing separate deployments

Noting that the scheduling spread strategy is centrally configured using the SelectorSpreadPriority, can you override the scheduling behavior on a per service or pod basis?